|Home | About | Journals | Submit | Contact Us | Français|
Estrogen signaling in auditory and vestibular sensory epithelia is a newly emerging focus propelled by the role of estrogen signaling in many other proliferative systems. Understanding the pathways with which estrogen interacts can provide a means to identify how estrogen may modulate proliferative signaling in inner ear sensory epithelia. Reviewed herein are two signaling families, EGF and TGFβ. Both pathways are involved in regulating proliferation of supporting cells in mature vestibular sensory epithelia and have well characterized interactions with estrogen signaling in other systems. It is becoming increasingly clear that elucidating the complexity of signaling in regeneration will be necessary for development of therapeutics that can initiate regeneration and prevent progression to a pathogenic state.
Hearing loss affects over 31 million Americans and this number is expected to grow substantially as the population ages (Pleis and Lethbridge-Cejku, 2006). Hearing deficits can lead to reduced physical and psychological well-being, depression, and isolation resulting in serious social and economic consequences (Liu and Yan, 2007; Moeller, 2007; Moeller et al., 2007). While hearing loss has a multifactor etiology, death of sensory hair cells is a common denominator in many forms of hearing impairment. Elucidation of signaling pathways involved in hair cell regeneration will help to provide tools for stimulating new hair cell production in the mature cochlea. Described herein is the current understanding of one such pathway, the estrogen signaling pathway, within the context of regeneration of mature auditory and vestibular sensory epithelia.
Originally characterized for its role in reproduction and gender differences, estrogen is emerging as a key player in stem cell differentiation, neuronal protection and regeneration research (Cao and Zhang, 2007; Gottfried-Blackmore et al., 2007). Discovery of these new roles has heightened efforts to delineate estrogen’s signaling mechanisms and determine how estrogen interacts with other signaling pathways. Estrogen initiates signaling through both membrane and nuclear receptors (Nilsson et al., 2001; Barkhem et al., 2004; Cheskis et al., 2007) and has a diverse repertoire of intracellular regulatory roles (Coleman and Smith, 2001; Kelly and Levin, 2001). Localized synthesis of estrogen creates niches of spatially restricted activity that contribute to its contextual signaling specificity (Chen et al., 2005; Lambard et al., 2005; Cornil et al., 2006; Leary and Dowsett, 2006; Ohnemus et al., 2006) and varied mechanisms by which estrogen can initiate signaling add to the complexity of the roles for estrogen in biological systems.
Estrogens are hormonally active molecules synthesized from androgens in situ by aromatase-mediated activity. The most abundant form of estrogen is 17-β-estradiol, which can be converted to estrone and, minimally, to estriol. The estrone and estriol metabolites have stronger affinities for estrogen receptors, but are weaker agonists than 17-β-estradiol (Kuiper et al., 1997). Estrone and estriol have limited, tissue-specific roles in signaling (Ding and Zhu, 2008; Watson et al., 2008), whereas 17-β-estradiol (hereafter “estrogen”) is the best-understood ligand. Estrogen is a lipid-soluble hormone that diffuses readily through the plasma membrane to bind to the nuclear receptors, estrogen receptor-α (ERα) and estrogen receptor-β (ERβ). ERα and ERβ are characterized by highly conserved DNA-binding (97%) and ligand-binding domains (60%). In contrast, the amino-terminal domain of the receptors share only 18% sequence homology and vary in length. This latter region is thought to be partially responsible for the signaling specificity of the ER isoforms (Hall et al., 2001; Heldring et al., 2007). In humans, there are three identified ERα isoforms (ERα1-3) and five ERβ isoforms (ERβ1-5). ERα isoforms 2 and 3 have not been identified outside cancer cell lines, leaving some speculation as to the existence of these splice variants in vivo. Splice variants of ERβ are expressed differentially in tissues and can modulate ERα signaling (Ogawa et al., 1998; Matthews and Gustafsson, 2003; Ramsey et al., 2004; Leung et al., 2006). For example, the ERβ2 isoform can inhibit ERα-dependent transcription when both receptors are expressed in the same cell (Matthews and Gustafsson, 2003). Beyond ERβ2, little else is known about the functions of other ERβ isotypes. For the purpose of this review, ‘ERα’ and ‘ERβ’ will refer to generalized receptor functions rather than tissue-specific isoforms.
Nuclear ERα and ERβ can be activated in a ligand-dependent and ligand-independent manner (Fig. 1). In ligand-dependent activation, heterodimerized ligand-bound receptors elicit direct genomic effects by binding an estrogen response element (ERE) in the promoter region of target genes. A conservative estimate of 600 possible target genes was reported from a screen for genes responsive to estrogen in the human genome, illustrating a wide array of genes potentially regulated by estrogen. Of these genes, many had putative EREs in their promoter regions for direct regulation of transcription by estrogen (Tang et al., 2004). However, nearly 35% of genes responsive to changes in estrogen are modulated through indirect regulation of transcription, where the receptor dimers bind and regulate other transcription factors (Marino et al., 2006). Direct and indirect activation of genes responsive to estrogen relies on recruitment of co-factors specific to the cellular signaling context. Non-genomic effects of estrogen signaling that modulate cytoplasmic or cell membrane-bound regulatory proforms proteins, independent of gene transcription or protein synthesis, can also be activated in a ligand-dependent manner.
In ligand-independent ER activation, activation of kinases by other signaling pathways results in phosphorylation and activation of ERs. The consequences of ER activation by phosphorylation are similar to that of ligand-mediated activation, with both genomic, and non-genomic functions (reviewed in Weigel and Zhang, 1998; Auger, 2001; Coleman and Smith, 2001; Osborne et al., 2001; Flint et al., 2002; Hart and Davie, 2002; Cvoro et al., 2006). Ligand-independent activation of ERs has been described for growth factors (Pietras et al., 1995; El-Tanani and Green, 1997; Marquez et al., 2001; Ma et al., 2007; Berno et al., 2008; Sinkevicius et al., 2008), dopaminergic agonists (Power et al., 1991; Blaustein, 2004; Cheskis et al., 2007), peptide hormones (e.g., GnRH) (Demay et al., 2001; Pak et al., 2006; Morales et al., 2007), cyclin-dependent kinase (Lecanda et al., 2007), protein kinase A, protein kinase C (Chen et al., 1999; Longo et al., 2006; Al-Dhaheri and Rowan, 2007) and other activators of cell signaling pathways (e.g., MAPK) (Kato et al., 1995; Bunone et al., 1996).
Although best characterized as nuclear receptors, limited expression of both ERα and ERβ occurs at the cell membrane. Signaling initiated by these membrane receptors is thought to contribute to non-genomic activity of estrogen by altering growth factor signaling and contributes to genomic functions by translocating as a ligand–receptor complex to the nucleus for ERE binding (Pappas et al., 1995; Levin, 1999; Clarke et al., 2000). Another membrane estrogen receptor is the recently discovered G-protein-coupled receptor, GPR30. Identification of this receptor helped solve the long-standing mystery of how the rapid non-genomic second messenger signaling of estrogen was mediated (Carmeci et al., 1997). GPR30 binds β-estradiol with the same affinity as ERα (Revankar et al., 2005). Transfection of GPR30 into cell lines and activation with estrogen, alters adenylyl cyclase activity (Thomas et al., 2005) and has been linked to transactivation of epidermal growth factor (EGF) receptors (Filardo, 2002). However, many uncertainties remain regarding the functions of this newly discovered estrogen receptor. Such as the potential interplay between GPR30 signaling and the traditional nuclear ERs, and details of the signaling mechanisms both at the membrane (Filardo and Thomas, 2005; Thomas and Dong, 2006; Watson et al., 2006; Filardo et al., 2007) and intracellularly (Matsuda et al., 2008; Teng et al., 2008; Wang et al., 2008).
Stem cells maintain the ability to divide and differentiate into specialized cells. Both human and mouse embryonic stem cells are estrogen-responsive, as are multiple types of stem-like progenitor cells (e.g., hematopoietic stem cells, neural stem cells), prompting a focus on estrogen signaling in stem cell differentiation and regeneration research (Korach, 1994; Son et al., 2005; Hong et al., 2006; Gilliver et al., 2007; Hong et al., 2007). Estrogen treatment in a mouse embryonic stem cell line promoted cell proliferation and an up-regulation of both ERα and ERβ, suggesting a direct effect of estrogen signaling on embryonic stem cell cycle regulation (Han et al., 2006). Estrogen increased proliferation of both glial cells and neural stem cells in neural cultures from embryonic rat (Brannvall et al., 2002). Similarly, estrogen increased proliferation of neuronal cultures taken from both male or female chicks (Cao and Zhang, 2007), suggesting that localized proliferative effects of estrogen are gender-independent. In addition, there are significant data regarding estrogen effects on proliferation in mature systems. For example, estrogen treatment after carotid artery injury in mature rats was found to promote both proliferation and migration of endothelial cells. These effects are partially attributable to an estrogen-dependent increase in vascular epidermal growth factor (VEGF), which promotes endothelial cell proliferation and re-endothelialization in vivo (Shifren et al., 1996; Concina et al., 2000; Kim and Levin, 2006). Although not correlated with estrogen activity, increased cochlear VEGF expression has been observed in the stria vascularis and in spiral ganglion cells of noise-exposed animals (Picciotti et al., 2006). Contributions of estrogen signaling to repair and recovery of mature epithelia by regulating proliferation has also been identified in pancreatic β-cells, cartilage, uterine epithelium and hippocampal neurons (Moraghan et al., 1996; Tanapat et al., 1999; Nephew et al., 2000; Talwar et al., 2006; Cano et al., 2008). Evidence is accumulating that the mechanisms for estrogen signaling in repair and recovery of other systems may extend to the mature inner ear.
Clinical hearing research for a role of estrogen in hearing has to contend with confounding factors such as presbycusis, differences in occupational noise exposure and combinatorial hormone replacement therapies. While these studies have not identified a direct role for estrogen in hearing, they do implicate estrogen in hearing function (Jonsson et al., 1998; Hultcrantz et al., 2006; Baraldi Gdos et al., 2007; Dreisbach et al., 2007). Men have consistent estrogen production throughout life, whereas women have steep cyclical changes in estrogen production, depending on the phase of the menstrual cycle, that ranges from concentrations equal to men to those 10-fold higher. After menopause estrogen production in ovaries ceases, and serum concentrations of estrogen for both men and women are essentially the same (reviewed in Cornil et al., 2006). This change provides a unique opportunity to study effects of estrogen on hearing loss by investigating changes in hearing pre- and post-menopause, the effects of estrogen in hormone replacement therapy or by comparing gender differences pre- and post-menopause with age-matched male counterparts. A study comparing mean air-conduction thresholds in post-menopausal women who were receiving either estrogen therapy alone, combined estrogen and progesterone therapy, or no hormone replacement therapy for an average duration of 4 years, showed a 10 dB shift at low frequencies in women receiving estrogen alone and 5 dB shift in women on combinatorial therapy, concluding that estrogen may slow hearing loss in aging women (Kilicdag et al., 2004). Another expansive epidemiological study correlated low serum estrogen levels in post-menopausal women with significant hearing loss when compared to post-menopausal women with no hearing loss. Hearing loss in this study was defined as a 40 dB hearing level at 1000 and 2000 Hz in one ear or both ears. Serum estrogen was 8.4+/−3.8 pg/ml for patients with hearing loss versus 9.9+/−8.7 pg/ml in women with normal hearing (Kim et al., 2002), supporting a role for estrogen in hearing protection. These findings were supported by another study showing an improvement in hearing threshold levels at low frequencies by 3–5 dB in post-menopausal women who received any hormone replacement therapy (HRT; estrogen alone or estrogen with progesterone). Further analysis showed post-menopausal women who were not on HRT also had a poorer hearing at low frequencies (3–5 dB loss) when compared to pre- and peri-menopausal women (Hederstierna et al., 2007). However, other studies show no effect of HRT on hearing protection (reviewed in Hultcrantz et al., 2006), which may be due in part to the antagonizing effects of progesterone and estrogen in vivo (Guimaraes et al., 2006) or the observation that protection from hearing loss may have only been seen in patients that suffered significant acoustic insult. Estrogen has been implicated in establishing gender differences in hearing during development prior to the onset of menses and ovarian estrogen production (McFadden, 2002; Guimaraes et al., 2004). Until recently much of the ongoing cell signaling research did not include gender as a variable, which may have masked possible estrogen effects relevant to normal auditory physiology.
The above clinical data demonstrate a positive effect of estrogen on hearing when at physiological concentrations. However, there have been reports of negative effects of estrogen on hearing. Several case studies describe reversible and irreversible hearing loss associated with oral contraceptive use (Okulicz, 1978; Hanna, 1986) and sudden temporary hearing loss associated with HRT (Strachan, 1996). These deficits in hearing have been attributed to thrombosis in the short term and/or otosclerotic bone loss in the inner ear over the long term. Because interplay between estrogen and other hormones contributes to the progression of these disorders, it is difficult to decipher the specific role of estrogen signaling in these instances. In guinea pig, long-term elevated non-physiological concentrations of estrogen resulted in severe hearing loss in some animals that was attributable to otosclerosis. Histological examination of those animals also revealed a thinning of the stria vascularis and loss of stereociliary bundles in the middle and apical turns of the cochlea (Horner et al., 2007).
In addition to the clinical data supporting a physiological role for estrogen in hearing, there is molecular evidence that estrogen may have a direct role in maintenance or repair of inner ear sensory epithelia. ERα is present during inner ear organogenesis in mouse (Sajan et al., 2007), and ERs are expressed in normal adult rat and mouse inner ear sensory epithelia and non-sensory epithelia (Stenberg et al., 1999; Meltser et al., 2008). A microarray study by Hawkins et al. (2007) detected significant up-regulation of estrogen receptors after either aminoglycoside antibiotic or laser damage of mature chicken auditory and vestibular epithelia. This finding is significant because avian auditory epithelia can regenerate after damage, unlike the mammalian auditory sensory epithelia known as the organ of Corti (OC). Therefore, the up-regulation of ERs seen in chicken suggests estrogen signaling has a role in repair of inner ear epithelia. A recent study by Meltser et al. (2008) was the first study to directly implicate estrogen signaling in recovery of the mammalian OC. This study used temporary threshold shifts to show reduced recovery of hearing thresholds after moderate acoustic trauma in knock-out mice lacking ERα, ERβ and aromatase, the enzyme responsible for localized estrogen synthesis. This deficiency was partially restored in the aromatase knock-out mice by treatment with a highly selective ERβ agonist, bypassing lack of localized E2 production and stimulating ERβ directly. Direct activation of ERα did not affect recovery rates in aromatase knock-out mice and these effects were not found to be gender specific (Meltser et al., 2008). Together, these data implicate estrogen as one of the pathways active in recovery of inner ear epithelia after acoustic damage.
The sensorineural hair cells are surrounded and sustained by non-sensory supporting cells. There are two mechanisms by which hair cells are replaced: (1) mitotic proliferation, where the supporting cells divide and a daughter cell differentiates into a hair cell (Corwin and Cotanche, 1988; Ryals and Rubel, 1988; Girod et al., 1989; Presson and Popper, 1990) and (2) direct transdifferentiation, a non-mitotic process in which supporting cells directly transform into hair cells without cell division (Raphael et al., 1994; Adler and Raphael, 1996; Roberson et al., 2004). Although it is possible to induce replacement of hair cells via both pathways, direct transdifferentiation may deplete the supporting cells necessary for healthy hair cell function and survival (Izumikawa et al., 2005; Cafaro et al., 2007). In mammals, the OC loses its ability to create new hair cells in utero (Chen and Segil, 1999) and the vestibular epithelia (saccule and utricle) cease hair cell production shortly after birth (Ruben, 1967). However, despite apparent quiescence, adult mammalian vestibular tissues maintain limited regenerative capabilities (Warchol et al., 1993; Lambert, 1994; Rubel et al., 1995; Li and Forge, 1997; Zheng and Gao, 1997; Zheng et al., 1997; Kuntz and Oesterle, 1998; Ogata et al., 1999; Li et al., 2003; Oesterle et al., 2003; Martinez-Monedero et al., 2007; Oshima et al., 2007) that can be stimulated by treatment with growth factors (Lambert, 1994; Yamashita and Oesterle, 1995; Corwin et al., 1996; Zheng et al., 1997; Kuntz and Oesterle, 1998; Oesterle et al., 2003). In contrast, agents to stimulate proliferation in the completely quiescent mature OC remain elusive (Sobkowicz et al., 1992; Roberson and Rubel, 1994; Chardin and Romand, 1995; Sobkowicz et al., 1996; Sobkowicz et al., 1997; Zine and de Ribaupierre, 1998). Ideal methods for restoring lost hair cells would replace hair cells while maintaining supporting cell populations. Promising approaches include transplantation of stem cells, induction of stem-like properties in existing cells, activating transdifferentiation, and stimulating proliferation of existing adult stem cells (reviewed in Collado et al., 2008). Recent studies in the neonatal mouse inner ear have identified a stem cell population in all sensory epithelia and demonstrate that vestibular tissues retain cells with stem-like capabilities throughout adulthood, further suggesting proliferative potential in the mature mammalian ear (Li et al., 2003; Oshima et al., 2007).
Non-pathological proliferation of the inner ear sensory epithelia can be categorized into two processes: developmental proliferation and regenerative proliferation. The extent to which these processes overlap is under investigation. Much of the current work to identify regulators of proliferation has focused on factors that affect proliferation during development or inhibit processes associated with maturation. Recapitulating development to replace hair cells lost to damage has had some limited success. One example shows disruption of the cell cycle control regulator, p27kip, promoted ongoing proliferation of supporting cells and ectopic expression of cells in the neonatal and adult organ of Corti (Chen and Segil, 1999; Lowenheim et al., 1999). However, the temporal and concurrent signaling necessary to coordinate regulation of p27kip and appropriate localization of replacement hair cells is ongoing. Another more direct approach to replace hair cells is to stimulate native auditory sensory epithelia repair/recovery pathways that can promote division of existing progenitor cells. Due to the complex array of signaling likely to be involved, this is apt to be accomplished through coordinated signaling that will simultaneously relieve inhibition while promoting proliferation of mature inner ear epithelia. Although cross-talk by estrogen signaling pathways in inner ear is a new focus, in many other tissues, estrogen is known to intimately interact with multiple signaling pathways (Imamov et al., 2005; Asarian, 2006; Bodo and Rissman, 2006; Cheskis et al., 2007; Heldring et al., 2007). Hence, we will focus here on the interaction of estrogen with growth factor signaling pathways that are known to be important for regulation of proliferation in the inner ear sensory epithelia. This may serve to provide insights into possible targets for molecular cross-talk between these pathways.
Homeostasis is the major energy investment in mature biological systems. Identification of factors that can release homeostatic, controls acting in the OC will be important for activating regenerative proliferation. However, such factors will only be therapeutically effective if repair can be stimulated without causing a novel pathology such as unregulated cell growth or oncogenesis. Therapies need to be developed that release inhibition, promote repair and then re-instate maintenance pathways at the proper time to restore tissues to the homeostatic state. This can be accomplished after better understanding the pathways, cross-talk and temporal organization of the active maintenance and repair signaling networks. Two families of growth factors, the epidermal growth factor (EGF) superfamily and the transforming growth factor β (TGFβ) superfamily, have shown promise in shifting adult mammalian inner sensory epithelia away from active maintenance, or quiescence, and into a modest proliferative state. Importantly, and as will be discussed later, both of these pathways interact with estrogen signaling in other systems.
The EGF receptors are a family of tyrosine kinase receptors that recognize over eleven different ligands (Table 1). There are four classes of EGF receptors, ErbB1-4, which can homo- or heterodimerize. Ligand-binding of ErbB receptors results in post-receptor regulation of a variety of signaling cascades (Riese and Stern, 1998; Yarden and Sliwkowski, 2001; Linggi and Carpenter, 2006). The pathway(s) activated by ErbB signaling are determined by the particular ligand-receptor pair and by the other signaling pathways that are active prior and coincident to ErbB activation. ErbB signaling is well characterized for its contribution to oncogenesis and the cellular effects of activation include cell division, cell death, motility, and adhesion (reviewed in Yarden and Sliwkowski, 2001; and Linggi and Carpenter, 2006). Despite very diverse signaling roles in other tissues, four tested ErbB ligands have been shown to potentiate proliferation of inner ear vestibular sensory epithelia.
The ErbB ligands that potentiate proliferation in mammalian vestibular sensory epithelia include transforming growth factor-α (TGFα, with and without insulin), EGF (with insulin), glial growth factor 2 (GGF2) and heregulin (HRG) (Lambert, 1994; Yamashita and Oesterle, 1995; Zheng et al., 1999; Montcouquiol and Corwin, 2001; Malgrange et al., 2002; Hume et al., 2003; Gu et al., 2007). While all of these ligands elicit mitogenic effects in the vestibular sensory epithelia, their effects differ with maturity and in interactions of ligand with insulin signaling. Regarding age-dependent sensitivity of vestibular tissue to growth factors, GGF2 and HRG stimulate proliferation in neonatal, but not mature, vestibular sensory epithelia. GGF2 and HRG are ligands derived from alternate splicing of the neuregulin-1 gene (Carraway and Burden, 1995), and both modestly potentiate proliferation in cultured neonatal utricles. Increased supporting cell proliferation was seen in dissociated utricular sheet cultures taken from rats before postnatal day 10 and treated with GGF2 or HRG (Zheng et al., 1999; Montcouquiol and Corwin, 2001; Gu et al., 2007). Neonatal organotypic utricular cultures were also sensitive to treatment with HRG (Zheng et al., 1999; Hume et al., 2003). Unlike EGF or TGFα, vestibular sensory epithelia have a restricted temporal sensitivity to GGF2 and HRG. Neither GGF2 or HRG significantly affected proliferation in utricular cultures from 4–6 week old mice (Hume et al., 2003) or rats older than postnatal day 10 (Gu et al., 2007). The finding that treatment of sensory epithelia with GGF2 or HRG has effects limited to young tissue supports the idea that before postnatal day 10, the signaling in rodent utricle may be continuing to mature. This loss of sensitivity to growth factors with continued maturity of the organ, at a stage when vestibular organs are thought to be fully developed (4–6 weeks), suggests a delayed onset of inhibitory regulation in the inner ear. These differences in responsiveness of developing versus mature tissues highlight the need to better understand active homeostatic mechanisms in mature tissues.
Regarding interactions of ErbB ligand with insulin signaling, insulin is a hormone involved in energy homeostasis that is known to modulate growth factor activity in many systems including inner ear sensory epithelia. In organotypic utricular cultures from neonatal mice, combined treatment of GGF2 and insulin increased proliferation by nearly 30% over either factor alone (Gu et al., 2007). In undamaged utricular cultures from mature mice, mitogenic effects of TGFα were potentiated by, but not dependent on, insulin, suggesting that cross-talk between TGFα and insulin signaling is potentially involved in the regenerative response (Yamashita and Oesterle, 1995; Kuntz and Oesterle, 1998). Co-regulation with insulin is also seen in organotypic cultures of mature mouse utricles treated with EGF and insulin concomitantly. Treatments showed a dose-dependent increase in supporting cell proliferation, however unlike TGFα, neither EGF or insulin alone significantly affected proliferation (Yamashita and Oesterle, 1995).
The mechanisms by which the erbB ligands and their co-regulation by insulin affect proliferation in inner ear epithelia are still unknown. These signaling pathways are sites for interaction with estrogen signaling in other regenerative systems, and it is likely that cross-talk between estrogen and ErbB signaling occurs in the ear as well. There are multiple examples of mutual regulation between estrogen and other cell signaling pathways. We describe three in detail below: interactions between estrogen and EGF, TGFα, and TGFβ.
One of the first characterized roles for the estrogen receptor GPR30 was the stimulation of EGF production (Filardo, 2002; Filardo et al., 2002; Filardo and Thomas, 2005; Filardo et al., 2008). Estrogen binding of GPR30 activates MAPK signaling cascades via adenylyl cyclase causing the release of heparin-bound EGF from the plasma membrane (Filardo et al., 2000). Despite the signaling role of GPR30 being primarily deduced using cancer cell lines, the role for GPR30 in EGF secretion and proliferation is likely to be relevant to non-pathogenic tissues. In some contexts, estrogen signaling activates EGF and, conversely, EGF can regulate estrogen signaling. For example, EGF initiates ligand-independent activation of estrogen signaling in mouse uterine epithelium, a mature epithelium that undergoes continued cyclical proliferation. This effect is not seen in the ERα knock-out mouse (Hewitt et al., 2005), suggesting that EGF activation is ERα-dependent rather than GPR30 receptor-dependent. Recent data from a knock-in mouse expressing a non-ligand-binding ERα, confirmed that ligand-independent activation of ERα by EGF promotes proliferation of estrogen-sensitive uterine epithelia (Sinkevicius et al., 2008). In inner ear vestibular sensory epithelia, EGF required insulin to potentiate proliferation. While not established in inner ear sensory epithelia, it is possible that EGF may increase estrogen secretion, which, in turn, stimulates insulin signaling (Asarian, 2006; Roepke et al., 2008) making both factors available for potentiating proliferation. Another example of estrogen affecting insulin secretion in a mature tissue is found in pancreatic β-cells. Treatment with estrogen increased insulin synthesis and secretion however, had no effect on proliferation (Alonso-Magdalena et al., 2008). Studies in mice lacking ERα or aromatase (the enzyme required for localized estrogen synthesis) confirmed that estrogen-induced insulin secretion in β-cells is an ERα mediated effect (Barros et al., 2006; Alonso-Magdalena et al., 2008). If we interpret these data to indicate a possible mechanism for the insulin/EGF effects shown in mouse utricle, a model for potentiating supporting cell proliferation is suggested in which the addition of exogenous insulin bypasses ER-stimulated insulin release to promote proliferation of inner ear sensory epithelia. While there are many potential mechanisms for cross-talk between these factors, it is intriguing to consider that temporally coordinated actions of EGF and estrogen could be elements required for regenerative proliferation of the inner ear sensory epithelia.
Another example of a possible mechanism for co-regulation in the inner ear is found in the mutual regulation of TGFα and estrogen in neuroblastoma cell lines. The presence or absence of ERα signaling pathways can act as a molecular switch, modifying the cellular response to TGFα. In ERα-positive cells, TGFα treatment blocked proliferation and promoted differentiation. When ERα was absent, TGFα treatment resulted in mitotic proliferation (Vyhlidal et al., 2000). Modulation of TGFα by ERα occurs via signal transducer and activation of transcription three (STAT3) signaling proteins (Ciana et al., 2003). Because estrogen modulation of TGFα function is strongly implicated in oncogenesis (Junier, 2000), it is difficult to decipher its normal physiological role. However, these studies suggest that investigation into TGFα signaling in auditory systems when ERα is present, may be useful in understanding how cell fate is determined.
TGFα, EGF and other growth factors can activate receptors and transcription factors common to both ER and ErbB signaling (Nagashima et al., 2008). Possible targets for manipulation of supporting cell proliferation in the auditory and vestibular tissues by combined ER and growth factor signaling include the Ras-activated/mitogen-activated protein kinases (MAPKs) pathway, the STAT pathway, the phospholipase C-γ pathway, and the Akt-activated/phophotidylinositol (PtdIns) 3-kinase pathway.
The TGFβ superfamily of ligands may also be a target for estrogen signaling in the inner ear sensory epithelia. The TGFβ family includes over 35 secreted cytokines including TGFβs, activins, bone morphogenic proteins (BMPs), growth differentiation factors (GDFs) and nodal (Table 2). All of the cytokines have tissue-, cell- and context-specific effects on cell growth and differentiation. TGFβ receptors are classified as type I or type II based on sequence homology. There are seven identified type I receptors and five type II receptors that comprise over 10 characterized receptor dimer combinations (de Caestecker, 2004). The ligands that signal through these receptor dimers have varying specificities. Some ligands bind to a single receptor pair, while others are promiscuous and can bind multiple receptor pairs with differing affinities. Activation of heterodimerized receptors results in phosphorylation of one or more of seven receptor associated Smad proteins. These include inhibitory Smads (Smad6/7) and receptor Smads, which are grouped into two families (Smad2/3 and Smad1/5/8) based on sequence homology. Once phosphorylated, the receptor-Smads or inhibitory-Smads associate with cytoplasmic Smad4, which facilitates translocation of the Smad complex to the nucleus where transcription is initiated or, in the case of Smad6/7, inhibited (Xu, 2006).
TGFβ ligands regulate progenitor cell proliferation in mature mammalian tissues predominantly by inhibiting proliferation but can have mitogenic effects (Li et al., 1998; Gandrillon et al., 1999; Pirskanen et al., 2000; McCroskery et al., 2003; Kawase et al., 2004). There are examples of differential effects of TGFβ ligands on cell proliferation in inner ear sensory epithelia. TGFβ ligands inhibited supporting cell proliferation in utricular sheet cultures from neonatal mice, (Zheng et al., 1997). However, activin, another ligand in the TGFβ family, potentiates auditory supporting cell proliferation in mature chicken cochlear ducts (McCullar et al., unpublished data). Multiple studies have demonstrated mutual regulation of the TGFβ superfamily and estrogen signaling (Kipp et al., 2007a,b). In breast cancer cells, activin represses estrogen-dependent transcription and conversely, estrogen treatment of these cells decreased activin-dependent transcription (Burdette and Woodruff, 2007). Interactions between these signaling pathways are not limited to cancerous tissues. In mature, normal uterine epithelium an environmental estrogen analog (bisphenol A) decreased expression of the type II activin receptor, Acvr2b, and its ligand, BMP7. Concurrent with down-regulation of Acvr2b and BMP7, there was a significant decrease in apoptotic cells, suggesting that estrogen regulation of activin signaling may affect cell survival. This effect was not seen in estrogen-treated testis or ovary, illustrating the tissue-specific context of the interaction (Kusumegi et al., 2004). Moreover, in normal mouse ovary, estrogen treatment specifically reduced transcription of activin. This transcriptional control on Acvr2a/b signaling by estrogen is mediated through the downstream activin signaling protein, Smad3 (Kipp et al., 2007a,b). However, regulation of Smad 3 is not due to direct binding or repression of Smad3 transcription by estrogen (Cherlet and Murphy, 2007), suggesting that estrogen may be interacting with other proteins in the activin signaling pathway as well. Together, these studies suggest that estrogen affects activin signaling both via altering transcription of ligand and by indirectly regulating the activin signaling protein Smad3. Other demonstrations of TGFβ cytokines interacting with estrogen signaling were shown in studies of TGFβ in endometrial renewal, wound healing and neural proliferation (Ashcroft et al., 1997; Ashcroft et al., 2003; Kanda and Watanabe, 2005; Buck and Knabbe, 2006; Elliot et al., 2006; Malek et al., 2006; Lecanda et al., 2007; Gargett et al., 2008; Kleuser et al., 2008) and in studies of BMP function in bone metabolism and stem cell proliferation (Yamamoto et al., 2002; Paez-Pereda et al., 2003; Zhou et al., 2003; Zhang et al., 2005; Lau et al., 2006).
We are just beginning to delineate factors that can restrict or stimulate supporting cell proliferation and hair cell replacement in mature inner ear sensory epithelia. Identification of ErbB ligands that affect proliferation of supporting cells in mature vestibular sensory epithelia is a promising start, but is limited to the vestibular sensory epithelia and the effects are small in magnitude. There are many questions that remain to be answered including: What factors are inhibiting cell proliferation in mature inner ear tissues? When do pathways necessary for maintaining supporting cell quiescence become active? What is the sequence of signaling necessary to take a supporting cell from quiescence into mitotic proliferation and differentiation and then return it to quiescence? How many cell-signaling pathways are involved? Work has already begun to collect large databases of genes responsive to injury in both mouse and chicken (e.g., Hawkins et al., 2007; Sajan et al., 2007). These data sets offer valuable clues for identifying new proteins involved in regulation of regeneration pathways. As more signaling pathways relevant to regeneration are discovered, the coregulation that controls homeostasis and regeneration can begin to be worked out. Estrogen signaling is of particular interest as a co-regulator due to known effects on proliferation in other tissues, its implication in survival and/or repair of inner ear, and its overlap with ErbB and TGFβ signaling pathways.
The authors’ research is supported by NIH NIDCD Grant DC005361, NIDCD P30 Grant DC04661, NICHHD P30 Grant HD002274and NIDCD DC003944. The authors would like to thank Drs. Monica Skinner, Allison Coffin, Kelly Owens, Yuan Wang, Nicole Schmitt and Martha Port for thoughtful comments and reading of this manuscript.